Cardiac myosin binding protein C regulates postnatal myocyte cytokinesis

Significance

We demonstrate that hearts lacking the sarcomere protein cardiac myosin binding protein C (MYBPC) undergo altered development due to an extra round of postnatal cell division. Normal cardiac myocytes replicate rapidly during fetal life, undergo a final round of cell division shortly after birth, cease dividing, and increase in cell size during prepubescent life. MYBPC has an unexpected function—inhibition of myocyte cytokinesis. MYBPC-deficient myocytes undergo an additional round of cytokinesis, resulting in increased numbers of myocytes and a greater proportion of mononuclear myocytes in neonatal hearts. Our findings provide insights into the mechanisms of dilated cardiomyopathy caused by homozygous mutations that reduce MYBPC levels.

Abstract

Homozygous cardiac myosin binding protein C-deficient (Mybpc^t/t) mice develop dramatic cardiac dilation shortly after birth; heart size increases almost twofold. We have investigated the mechanism of cardiac enlargement in these hearts. Throughout embryogenesis myocytes undergo cell division while maintaining the capacity to pump blood by rapidly disassembling and reforming myofibrillar components of the sarcomere throughout cell cycle progression. Shortly after birth, myocyte cell division ceases. Cardiac MYBPC is a thick filament protein that regulates sarcomere organization and rigidity. We demonstrate that many Mybpc^t/t myocytes undergo an additional round of cell division within 10 d postbirth compared with their wild-type counterparts, leading to increased numbers of mononuclear myocytes. Short-hairpin RNA knockdown of Mybpc3 mRNA in wild-type mice similarly extended the postnatal window of myocyte proliferation. However, adult Mybpc^t/t myocytes are unable to fully regenerate the myocardium after injury. MYBPC has unexpected inhibitory functions during postnatal myocyte cytokinesis and cell cycle progression. We suggest that human patients with homozygous MYBPC3-null mutations develop dilated cardiomyopathy, coupled with myocyte hyperplasia (increased cell number), as observed in Mybpc^t/t mice. Human patients, with heterozygous truncating MYBPC3 mutations, like mice with similar mutations, have hypertrophic cardiomyopathy. However, the mechanism leading to hypertrophic cardiomyopathy in heterozygous MYBPC3^+/− individuals is myocyte hypertrophy (increased cell size), whereas the mechanism leading to cardiac dilation in homozygous Mybpc3^−/− mice is primarily myocyte hyperplasia.

Dilated cardiomyopathy (DCM) leads to heart failure and is a leading cause of morbidity and mortality (1, 2). DCM is generally diagnosed as left ventricular (LV) dilation with associated reduction in cardiac contraction measured as impaired fractional shortening (3). Hearts from affected individuals frequently demonstrate myocyte elongation, myocyte death, and fibrosis, in addition to LV dilation. DCM results from a variety of environmental factors, such as viral infection and alcohol abuse, as well as from mutations in a number of genes including titin, lamin A/C, cardiac actin, cardiac myosin heavy chain, and phospholamban (reviewed in refs. 4–6). Whether all of these DCM-inducing factors activate the same or different cellular pathways to produce similar clinical features remains uncertain. The mechanisms by which mutations in the cardiac myosin binding protein C (MYBPC3) gene and other sarcomere protein genes lead to cardiac dilatation are under investigation.

MYBPC is a thick filament accessory protein component of the striated muscle sarcomere A band that constitutes 2–4% of the myofibril (discussed in ref. 7). Although there are four Mybpc genes in the mammalian genome, only cardiac Mybpc (Mybpc3) is expressed in embryonic, neonatal, and adult hearts (8, 9). Cardiac MYBPC interacts with at least four sarcomere components: myosin heavy chain, actin, myosin light chain 2, and titin (10–12). More than 400 cardiac MYBPC3 gene mutations have been identified in patients as a cause of hypertrophic cardiomyopathy (HCM), an autosomal dominant disorder resulting from defective sarcomeres (for reviews, see refs. 12, 13). Due to an ancient founder mutation, 4% of the population of India carries a truncating MYBPC3 mutation (14, 15). The majority of cardiac MYBPC3 mutations are predicted to encode truncated proteins that lack portions of either the carboxyl myosin and/or titin binding domains (7, 13). These truncating MYBPC3 mutations are thought to cause cardiac hypertrophy by inducing myocyte hypertrophy (increased cell size), rather than myocyte hyperplasia.

We and other researchers have created mice that carry a mutant cardiac Mybpc3 gene to create murine HCM models (16–18). Heterozygous mice, designated Mybpc^t/+, like humans bearing the same mutation, develop adult onset HCM. Homozygous MYBPC3 mutations are a much rarer cause of human DCM than autosomal dominant mutations in other sarcomere protein genes. However, homozygous Mybpc^t/t mice that express two mutant alleles and no wild-type cardiac Mybpc3 develop LV dilation by 3 d postbirth and have all of the features of DCM, including LV chamber dilation, albeit mildly impaired fractional shortening (16). Unlike most humans with DCM, homozygous mutant cardiac Mybpc^t/t mice have normal survival despite their cardiac disease. Other homozygous null cardiac Mybpc3 mice develop an identical phenotype (7, 17, 18). Hence, for the studies described here, we assume that the phenotype of the Mybpc^t/t mice is due to lack of MYBPC protein, rather than to small amounts of truncated protein. Recently, two groups have demonstrated that delivery of MYBPC to Mybpc3-null hearts restores cardiac function and morphology (19, 20). Here, we have begun to dissect the mechanism by which homozygous Mybpc^t/t hearts develop DCM.

Because Mybpc^t/t mice begin LV dilation within a few days postbirth (16), we hypothesized that this reflected abnormal development of neonatal myocytes. During fetal and early perinatal development in wild-type hearts, cardiomyocytes divide rapidly, producing hyperplastic cardiac growth (21). However, at 10 d postbirth, cardiomyocytes cease to divide and all subsequent increases in myocardial mass result from myocyte hypertrophy (22). Despite the importance of this phenomenon, little is known about the molecular basis for the transition from hyperplasic to hypertrophic-based myocardial growth. We hypothesized that abnormal cardiomyocyte growth, either hyperplastic or hypertrophic, in the perinatal period accounted for the LV dilation of Mybpc^t/t mouse hearts. To address this question, we have counted and measured cardiomyocytes from Mybpc^t/t and wild-type mice. We have also studied the consequences of reducing MYBPC levels by injecting Mybpc3-specific shRNA at birth. Neonatal cardiomyocytes lacking cardiac MYBPC, due to Mybpc3-specific shRNA knockdown, undergo an additional round of cytokinesis. We conclude that dramatic reductions in the amount of cardiac MYBPC leads to aberrant cell cycle regulation at the G1/S checkpoint, resulting in at least one extra round of myocyte division and DCM.

Results

Increased Numbers and Immaturity of Cardiac Myocytes in Mybpc^t/t Mice.

Cardiac tissues and myocytes from wild-type and Mybpc^t/t mice were studied from birth to postnatal day 35 (designated as P0–35). The hearts from P5 Mybpc^t/t mice had a significantly increased LV mass and LV/body weight ratio compared with wild-type mice (Fig. 1 A and B and Fig. S1A). Myocytes from P3 and P10 Mybpc^t/t mice were 20–30% wider than wild-type myocytes (Fig. S1 B and C), which we presume reflects myocyte immaturity (23) and/or abnormal spacing between parallel sarcomeres due to MYBPC deficiency (24). There was no significant difference in the length of myocytes from wild-type and Mybpc^t/t mice (Fig. S1D).

Fig. 1.

LV mass and myocyte nucleation in wild-type and Mybpc^t/t hearts. (A and B) Changes in LV mass (A) and LV weight/body weight ratios (B) of wild-type (filled circle) and Mybpc^t/t (open circle) mice from P0–20 (n = 5–10 neonates per time point). (C) Light micrographs of DAPI-stained isolated mononuclear (Upper panel) and binuclear (Lower panel) myocytes from 3-wk-old Mybpc^t/t mice visualized with bright field (BF) and fluorescent (DAPI) illumination. (Scale bar, 20 µm.) (D) Representative light micrographs of isolated LV myocytes from EdU-labeled Mybpc^t/t mice that are stained for nuclei (blue), troponin I (green), and EdU (red). A mononuclear myocyte (arrows) is indicated. (Scale bar, 20 µm.) (E) Distribution of mononuclear, binuclear, and multinuclear myocytes from 3-wk-old wild-type (gray), Mybpc^t/+(white box), and Mybpc^t/t (black) hearts. Data are presented from four mice per genotype, 200–300 cells per mouse (mean ± SD). All P values reflect comparison of wild-type and Mybpc^t/t hearts.

Fig. S1.

(A) DAPI-stained transverse sections of wild-type (Left) and Mybpc^t/t (Right) hearts from P3 (Upper) and P10 (Lower) mice. (Scale bar, 0.5 mm.) (B) Isolated ventricular myocytes from P3 wild-type (Top Left), P3 Mybpc^t/t (Top Right), P10 wild-type (Bottom Left), and P10 Mybpc^t/t (Bottom Right) fixed with 4% (wt/vol) paraformaldehyde and stained with propidium iodide (red) and phalloidin (green). (Scale bar, 50 µm.) (C) Width and (D) length of myocytes isolated from wild-type (gray) and Mybpc^t/t (black) in P3 and P10 mice (n = 13–18 cells, three animals per group).

As expected (22, 25, 26), nearly all wild-type adult LV myocytes were binuclear (Fig. 1 C–E). Analyses of younger mice confirmed this finding. At P21, ∼5% of wild-type and heterozygous Mybpc^t/+ LV myocytes were mononuclear and 90% were binuclear (Fig. 1E). However, significantly less Mybpc^t/t LV myocytes were binuclear (Fig. 1E; WT:Mybpc^t/+:Mybpc^t/t, 89.8%:89.7%:74.8%; P = 0.0003). Consistent with this observation was the observation that ∼threefold more Mybpc^t/t LV myocytes were mononuclear than either wild-type or heterozygous Mybpc^t/+ LV myocytes (Fig. 1E; WT:Mybpc^t/+:Mybpc^t/t, 5.8%:6.7%:18.3%; P < 4E-5). We defined myocyte numbers and nuclear morphology in hearts from 5-wk-old Mybpc^t/t and wild-type mice by immunohistochemical staining of 10 sections evenly distributed across the LV. Wheat germ agglutinin (WGA) was used to demarcate plasma membrane boundaries (Fig. 2 A and B), and nonmyocytes were excluded by size and absence of cardiac troponin I staining. In comparison with wild-type, Mybpc^t/t LV contained ∼40% more myocytes than wild-type LV (Fig. 2C; P = 0.006).

Fig. 2.

Increased numbers of myocytes in Mybpc^t/t compared with wild-type hearts. (A and B) Transverse heart sections stained with WGA at low (A; scale bar, 0.5 mm) and high (B; scale bar, 50 µm) magnification. (C) Quantification of myocytes from wild-type (gray) and Mybpc^t/t (black) mice. Data are presented from three mice per genotype, 10 sections per mouse (mean ± SD).

Previous studies have demonstrated that by P10 the majority of wild-type binuclear myocytes have exited the cell cycle (27). We hypothesized that the increased proportion of mononuclear to binuclear myocytes in Mybpc^t/t hearts reflected either premature cell cycle arrest before becoming binuclear or alternatively that mutant myocytes had continued cell cycle progression with cytokinesis—a delay in cell cycle exit that would account for greater myocyte number. To distinguish these models, we used three approaches to study cell cycle progression.

Mybpc^t/t Myocytes Have Delayed Cell Cycle Exit.

We injected the thymidine analog 5-ethynyl-2’-deoxyuridine (EdU) (28) into wild-type and Mybpc^t/t neonates at P1 through P5 and assessed DNA synthesis in isolated P14 myocytes (Fig. 1D). In comparison with wild type, there were significantly more EdU-positive mono- and binuclear myocytes from Mybpc^t/t mice (47.02 ± 2.50% vs. 29.35 ± 3.06%; P = 0.0003; four mice per genotype; 200–300 cells per mouse). To identify cells undergoing mitosis, we used an anti-histone H3 phosphorylation (pH3) antibody that detects pH3 of serine residue 10 during the mitotic (M) phase of the cell cycle (29). Immunohistochemistry analyses of pH3 at P3, P7, and P10 (Fig. 3 A and B) indicated almost twice as many myocytes undergoing mitosis in Mybpc^t/t than in wild-type hearts. Finally, we performed immunostaining for Aurora B kinase (Aurora B), which localizes to the central spindle during anaphase and in the midbody during cytokinesis (30). As cell cycle exit is arrested before cytokinesis during normal neonatal myocyte development, Aurora B kinase was rarely detected in the central spindle or midbody of wild-type myocytes. In contrast, P7 Mybpc^t/t myocytes had 15-fold more Aurora B staining, which was localized to the cytoplasmic bridge between daughter cells (Fig. 3 C and D). Based on the increase in EdU incorporation and pH3 and Aurora B immunostaining in Mybpc^t/t mice, we concluded that mutant myocytes have delayed cell cycle exit during neonatal life and undergo additional rounds of cell division. As a result, Mybpc^t/t mice had both more myocytes and more mononuclear myocytes than wild-type hearts.

Fig. 3.

pH3 and Aurora B expression in wild-type and Mybpc^t/t hearts. (A) Confocal micrographs of P7 wild-type and Mybpc^t/t LV sections stained with DAPI (blue), troponin I (green), and pH3 (red). (Scale bar, 20 µm.) (B) Quantification of pH3-positive myocytes in wild-type (gray) and Mybpc^t/t (black) hearts from P3, P7, and P10 mice. Data are presented from four mice per genotype, three sections per mouse (mean ± SD). (C) Confocal micrographs of cardiac sections from P7 wild-type and Mybpc^t/t mice stained with DAPI (blue), troponin I (green), and Aurora B (red). [Scale bar, (Upper) 20 µm and (Lower) 5 µm.] (D) Quantification of Aurora B located at the bridge of two daughter myocytes in wild-type (gray) and Mybpc^t/t (black) hearts from P7 mice. Data are presented from four mice per genotype, three sections per mouse (mean ± SD).

To further assess cell cycle regulation in neonatal (P10) Mypbc^t/t LV, we assessed expression of cell cycle-related genes by RNAseq (Table S1). One hundred and seven genes that are described as “cell division,” “cyclin-associated,” or “cell cycle” genes are expressed in the wild-type P10 LV. Expression of 70 cell cycle-related genes either increased or decreased (P < 0.001), whereas 38 of the cell cycle-related genes were expressed at similar levels in wild-type and Mypbc^t/t LV. Presumably these cell cycle-associated gene expression changes are related to changes in cell cycle exit found in neonatal Mypbc^t/t LV.

Table S1.

Cell cycle-associated gene expression at P10 in wild-type and Mypbc^t/t LV

Gene	Chr	Start	End	Description	WT	Mybpct/t	Fold	P value
Differentially expressed cells
Ccnyl1	chr1	64737786	64772216	Cyclin Y-like 1	9.9	5.4	1.84	4.E-08
Ccnj	chr19	40905768	40923060	Cyclin-J	13.7	7.7	1.78	6.E-10
Cdk13	chr13	17807795	17896931	Cell division protein kinase 13	24.3	14.5	1.68	9.E-14
Bod1	chr11	31565149	31571862	Biorientation of chromosomes in cell division	24.1	14.5	1.66	2.E-13
Bod1l	chr5	42178777	42235554	Biorientation of chromosomes in cell division	52.3	32.5	1.61	5.E-24
Cdk16	chrX	20265096	20277003	Cell division protein kinase 16	158.2	98.7	1.60	2.E-68
Ccni	chr5	93610958	93635521	Cyclin-I	99.6	64.1	1.56	6.E-39
Cdk12	chr11	98064618	98139845	Cell division protein kinase 12	23.5	15.6	1.51	3.E-09
Cntd1	chr11	101140325	101150015	Cyclin N-terminal domain-containing protein 1	7.7	5.1	1.50	7.E-04
Cdc42ep3	chr17	79733364	79754431	Cdc42 effector protein 3	17.8	12.0	1.48	7.E-07
Cdk2ap1	chr5	124795447	124804637	Cyclin-dependent kinase 2-associated protein 1	38.8	26.6	1.46	2.E-12
Cdc73	chr1	145454628	145549814	Parafibromin	8.9	6.1	1.45	7.E-04
Cdc42bpb	chr12	112531182	112615929	Serine/threonine-protein kinase MRCK beta	95.7	67.3	1.42	3.E-25
Dmtf1	chr5	9118867	9161776	Cyclin D binding myb-like transcription factor 1	11.9	8.4	1.41	3.E-04
Cdc5l	chr17	45528838	45570656	Cell division cycle 5-related protein	26.7	20.6	1.30	3.E-05
Cdc37l1	chr19	29064983	29101169	Hsp90 cochaperone Cdc37-like 1	55.7	43.1	1.29	4.E-09
Cdkn1c	chr7	150644243	150646955	Cyclin-dependent kinase inhibitor 1C	28.8	22.5	1.28	4.E-05
Ccny	chr18	9314041	9450148	Cyclin-Y	39.6	31.6	1.25	1.E-05
Cdc42bpa	chr1	181891219	182095733	Cell division cell 42 binding kinase alpha	41.3	33.4	1.23	3.E-05
Ccnd1	chr7	152115835	152125830	G1/S-specific cyclin-D1	84.1	68.5	1.23	5.E-09
Ccar1	chr10	62206675	62255116	Cell division cycle and apoptosis regulator	70.4	82.5	0.85	1.E-05
Ccnk	chr12	109417947	109441569	Cyclin-K	37.6	45.3	0.83	1.E-04
Cdc123	chr2	5715339	5766006	Cell division cycle protein 123 homolog	40.1	49.7	0.81	6.E-06
Ccnd2	chr6	127075726	127101066	G1/S-specific cyclin-D2	304.1	377.0	0.81	1.E-36
Cdc23	chr18	34790604	34811390	Cell division cycle protein 23 homolog	33.1	41.1	0.81	3.E-05
Cdkn1b	chr6	134870418	134875543	Cyclin-dependent kinase inhibitor 1B	63.1	78.8	0.80	3.E-09
Cdc42bpg	chr19	6306456	6325652	Serine/threonine-protein kinase MRCK gamma	19.1	24.2	0.79	6.E-04
Cdk7	chr13	101466979	101500897	Cell division protein kinase 7	26.6	33.9	0.78	2.E-05
Ccng2	chr5	93696598	93705257	Cyclin-G2	16.4	21.1	0.78	5.E-04
Ccna2	chr3	36463786	36470918	Cyclin-A2	14.0	18.2	0.77	8.E-04
Cdca4	chr12	114058445	114067634	Cell division cycle-associated protein 4	14.6	19.3	0.76	3.E-04
Cdc34	chr10	79144939	79151143	Ubiquitin-conjugating enzyme E2 R1	54.1	72.8	0.74	7.E-14
Cdc25b	chr2	131012686	131024233	M-phase inducer phosphatase 2	8.0	11.6	0.70	4.E-04
Ccnt1	chr15	98373641	98398714	Cyclin-T1	21.0	30.4	0.69	5.E-09
Cdk4	chr10	126500658	126504338	Cell division protein kinase 4	59.0	85.7	0.69	2.E-23
Cdk2ap2	chr19	4097350	4099017	Cyclin-dependent kinase 2-associated protein 2	14.8	21.5	0.69	7.E-07
Cdc37	chr9	20942984	20954350	Hsp90 cochaperone Cdc37	108.4	161.9	0.67	2.E-48
Cdc42ep1	chr15	78673076	78681332	Cdc42 effector protein 1	62.8	94.7	0.66	3.E-30
Cdk2	chr10	128134994	128142107	Cell division protein kinase 2	20.0	31.3	0.64	2.E-12
Cdc42ep4	chr11	113588163	113613129	Cdc42 effector protein 4	18.3	29.2	0.63	1.E-12
Ptgis	chr2	167028695	167066037	Prostacyclin synthase	17.4	28.2	0.62	8.E-13
Cdc27	chr11	104363842	104411934	Cell division cycle protein 27 homolog	27.7	46.9	0.59	3.E-23
Ccndbp1	chr2	120834143	120842648	Cyclin-D1–binding protein 1	21.0	36.4	0.58	1.E-19
Cks1b	chr3	89219393	89222213	Cyclin-dependent kinases regulatory subunit 1	17.7	31.0	0.57	1.E-17
Rmnd1chr10+	chr10	5914188	5943372	Required for meiotic nuclear division protein 1	16.1	29.0	0.55	9.E-18
Cdc16	chr8	13757689	13781882	SubName: Full, Cdc16 protein	24.0	43.4	0.55	5.E-26
Cdk6	chr5	3343892	3522225	Cell division protein kinase 6	6.6	11.9	0.55	3.E-08
Ccne2	chr4	11118500	11131926	G1/S-specific cyclin-E2	5.2	9.5	0.55	6.E-07
Cdkn2dchr9+	chr9	123836859	123839607	Cyclin-dependent kinase 4 inhibitor D	14.0	26.4	0.53	3.E-18
Cdkn2c	chr4	109333480	109339262	Cyclin-dependent kinase 4 inhibitor C	20.3	38.9	0.52	3.E-27
Cdkn2dchr9-	chr9	21092907	21095653	Cyclin-dependent kinase 4 inhibitor D	14.9	29.0	0.52	4.E-21
Cinp	chr12	112110819	112127355	Cyclin-dependent kinase 2-interacting protein	20.1	40.3	0.50	5.E-31
Cdca8	chr4	124595708	124614161	Borealin	7.7	16.4	0.47	6.E-15
Cdk1	chr10	68799382	68815660	Cell division protein kinase 1	15.6	33.4	0.47	8.E-30
Cdkn3	chr14	47380215	47391200	Cyclin-dependent kinase inhibitor 3	2.3	4.9	0.46	1.E-05
Cdc42se1	chr3	95032701	95040331	CDC42 small effector protein 1	33.4	73.3	0.46	1.E-66
Cks2	chr13	51740600	51746031	Cyclin-dependent kinase regulatory subunit 2	11.7	25.8	0.45	8.E-25
Cdca5	chr19	6085096	6091773	Cell division cycle associated 5	2.4	5.5	0.43	8.E-07
Mnd1	chr3	83891855	83959708	Meiotic nuclear division protein 1 homolog	1.4	3.2	0.43	2.E-04
Cdc42ep5	chr7	4102861	4116301	Cdc42 effector protein 5	12.1	30.2	0.40	3.E-35
Cdca2	chr14	68294410	68333898	Cell division cycle-associated protein 2	3.9	9.8	0.39	1.E-12
Cdca7l	chr12	119082333	119117179	Cell division cycle-associated 7-like protein	0.9	2.4	0.38	5.E-04
Cdkl1	chr12	70847834	70891694	Cyclin-dependent kinase-like 1	3.2	9.4	0.34	1.E-14
Cdc20	chr4	118105505	118109948	Cell division cycle protein 20 homolog	10.5	31.7	0.33	4.E-48
Cacybp	chr1	162133051	162142908	Calcyclin-binding protein	32.9	102.6	0.32	4.E-157
Cdkn1a	chr17	29227930	29237667	Cyclin-dependent kinase inhibitor 1	49.6	161.5	0.31	8.E-259
Cdc6	chr11	98769202	98785256	Cell division control protein 6 homolog	3.3	10.7	0.30	8.E-19
Ccnb2	chr9	70255495	70269361	G2/M-specific cyclin-B2	13.1	43.8	0.30	2.E-73
Cdca3	chr6	124780193	124783719	Cell division cycle-associated protein 3	9.9	33.3	0.30	2.E-56
Cdc25c	chr18	34892650	34911187	Cell division cycle 25 homolog C	3.6	13.4	0.27	2.E-26
Not differentially expressed cells
Cdk18	chr1	134010123	134036262	Cell division protein kinase 18	10.5	8.0	1.30	8.E-03
Cdk8chr5+	chr5	147042825	147114450	Cell division protein kinase 8	9.8	7.6	1.29	1.E-02
Cdc14b	chr13	64290579	64376296	Cell division cycle 14b	15.0	11.7	1.28	3.E-03
Ccnb1	chr13	101548693	101556441	G2/M-specific cyclin-B1	3.3	2.6	1.27	2.E-01
Cdk8chr10+	chr10	40069113	40203624	Cell division protein kinase 8	15.0	12.2	1.23	1.E-02
Cdk19	chr10	40059369	40203624	Cell division cycle 2-like 6	15.2	12.4	1.22	2.E-02
mKIAA0432	chr17	45544744	45570656	Cell division cycle 5-like (Schizosaccharomyces pombe)	19.5	16.4	1.19	2.E-02
Scaper	chr9	55397689	55785922	S-phase cyclin A-associated protein	13.1	11.3	1.16	9.E-02
Nucks1	chr1	133807034	133832898	Nuclear ubiquitous casein and cyclin-dependent	50.2	44.1	1.14	3.E-03
Cdc14a	chr3	115975470	116126950	Dual specificity protein phosphatase CDC14A	2.3	2.1	1.07	7.E-01
Cdc42	chr4	136875610	136913652	Cell division control protein 42 homolog	122.8	116.4	1.06	6.E-02
Fam58b	chr11	78564007	78565231	Cyclin-related protein FAM58A	8.3	8.0	1.03	7.E-01
Cdk17	chr10	92623620	92729818	Cell division protein kinase 17	26.4	25.9	1.02	7.E-01
Cdc25a	chr9	109778082	109796394	M-phase inducer phosphatase 1	8.0	7.9	1.02	8.E-01
Ccnjl	chr11	43342285	43400499	Cyclin-J–like protein	3.6	3.5	1.01	9.E-01
Ccnd3	chr17	47641999	47736637	G1/S-specific cyclin-D3	91.3	90.6	1.01	8.E-01
Cdk5	chr5	23924059	23929348	Cell division protein kinase 5	18.7	18.7	1.00	1.E+00
Ccng1	chr11	40562053	40568788	Cyclin-G1	191.9	193.1	0.99	8.E-01
Ccnf	chr17	24360176	24388354	Cyclin-F	4.9	5.0	0.99	1.E+00
Cdc42se2	chr11	54530916	54601205	CDC42 small effector protein 2	24.6	24.8	0.99	9.E-01
Cdk11b	chr4	154998977	155024041	Cell division cycle 2-like 1	66.5	67.8	0.98	6.E-01
Cdk9	chr2	32561301	32568304	Cell division protein kinase 9	16.6	17.2	0.97	7.E-01
Cdc26	chr4	62042609	62069657	Cell division cycle 26	18.4	19.0	0.96	6.E-01
Cdc42ep2	chr19	5916127	5924816	Cdc42 effector protein 2	36.7	38.8	0.95	3.E-01
Pigu	chr2	155103987	155183160	Cell division cycle protein 91-like 1	22.6	24.5	0.92	2.E-01
Cdc40	chr10	40552698	40602949	Pre-mRNA–processing factor 17	10.7	11.8	0.91	3.E-01
Mrrf	chr2	35991917	36045803	Ribosome-recycling factor, mitochondrial	35.7	39.4	0.91	5.E-02
Ccnh	chr13	85329081	85353328	Cyclin-H	17.3	19.5	0.89	1.E-01
Inca1	chr11	70501862	70513657	Inhibitor of CDK interacting with cyclin A1	1.9	2.3	0.82	4.E-01
Cdkl2	chr5	92435100	92472044	Cyclin-dependent kinase-like 2	2.4	3.0	0.81	3.E-01
Cdk10	chr8	125748740	125756156	Cell division protein kinase 10	12.1	14.8	0.81	2.E-02
Cdc7	chr5	107393340	107413450	Cell division cycle 7-related protein kinase	3.0	3.8	0.78	2.E-01
Rmnd1chr19+	chr19	12979720	12981475	Required for meiotic nuclear division protein 1	13.5	17.3	0.78	2.E-03
Cdcrel-1	chr16	18621903	18630031	Cell division cycle 1 homolog; aka septin 5	8.2	10.7	0.76	1.E-02
Cdk14	chr5	4803384	5380251	Cell division protein kinase 14	7.1	9.8	0.72	3.E-03
Cdkl3	chr11	51817722	51898169	Cyclin-dependent kinase-like 3	6.2	8.8	0.70	3.E-03
Cdc45	chr16	18780539	18812065	Cell division control protein 45 homolog	3.7	5.5	0.67	9.E-03
Cdca7	chr2	72314275	72324947	Cell division cycle-associated protein 7	3.2	5.3	0.60	1.E-03

Chr, chromosome; end, mm9 base pair position; fold, fold difference between WT and Mybpc^t/t gene expression at P10; P value, probability that the difference between Wt and Mybpc^t/t gene expression would occur by chance; start, mm9 base pair position. WT (wild-type) and Mybpc^t/t RNA expression of selected cell cycle genes is presented in reads per mapped reads.

Neonatal Depletion of Cardiac MYBPC Impairs Myocyte Maturation in Wild-Type Mice.

To confirm that altered myocyte maturation in Mybpc^t/t hearts was due to the myocyte-autonomous deficiency of MYBPC, we used the cardiotropic adeno-associated virus serotype 9 (AAV9) (31, 32) to deliver Mybpc3-specific shRNA to the neonatal heart. Two Mybpc3-specific shRNAs that targeted different regions of Mybpc3 RNA (Fig. S2A) were constructed with a colinear EGFP reporter gene downstream of the cardiac troponin T (cTnT) promoter, so as to exclude nonmyocyte expression (Fig. S2B).

Fig. S2.

AAV-mediated depletion of Mybpc3 in wild-type mice. (A) Location of 21-bp shRNA sequences targeting the Mypbc3 transcript. (B) Schematic representation of the AAV vector containing a colinear cTnT promoter and Mybpc3 shRNA-1 or shRNA-2.

AAV9-Mybpc3-shRNA was injected into the thoracic cavity of P1 wild-type neonates [5 × 10¹³ viral genomes (vg)/kg], and EGFP fluorescence was assessed to monitor shRNA expression. Cardiac EGFP fluorescence was detected 48 h after virus injection and continued for at least 5 mo (Fig. 4A) but was absent from all other organs (33). EGFP fluorescence was present in 60–80% of adult myocytes isolated from viral-transduced mice (Fig. 4B). Both of these shRNA constructs (denoted AAV9-Mybpc3-shRNA) had comparable efficiency in attenuating in vivo Mybpc3 RNA expression (Fig. 4C).

Fig. 4.

Depletion of Mybpc3 RNA by AAV-shRNA. (A) Hearts from three wild-type mouse hearts harvested 48 h after transduction with AAV9 lacking EGFP (center, arrow) or with AAV9-EFFP (two flanking center heart) visualized under fluorescent light. Right panel shows two hearts harvested 5 mo after AAV9-EFFP transduction. (Scale bar, 2 mm.) (B) Fluorescent microscopy of LV myocytes stained by DAPI (blue) and EGFP (green) isolated from wild-type mice transduced with AAV-LacZ shRNA (control, Upper panels) or AAV-Mybpc3 shRNA (Lower panels). DAPI was shown in red in merged images. (Scale bar, 50 µm.) (C) Quantitative real-time PCR analysis of LV Mybpc3 expression from mice transduced with AAV-Mybpc3 shRNA-1 or shRNA-2. Transcript levels were normalized to expression of the control AAV-LacZ shRNA. Data from three mice per group are presented (mean ± SD). (D) Distribution of mononuclear, binuclear, and multinuclear EGFP-expressing myocytes from 4-wk-old wild-type mice transduced (at P1) with AAV-LacZ shRNA (gray), AAV-Mybpc3 shRNA-1 (black), and AAV-Mybpc3 shRNA-2 (white). Data are presented from four mice per genotype, 200–300 cells per mouse (mean ± SD).

Four weeks after viral injection, EGFP-positive myocytes were isolated, and nuclear morphology was examined. Wild-type mice infected with a control shRNA (that targeted LacZ RNA; Materials and Methods) had ∼5% mono- and 90% binuclear myocytes (Fig. 4D). Among EGFP-positive myocytes, wild-type hearts injected with Mybpc3-specific shRNA had 5–6-fold more mononuclear myocytes than hearts injected with control shRNA (Fig. 4D). Mice infected with either of the two distinct Mybpc3-specific shRNA constructs showed similar results (Fig. 4D).

Myocyte Proliferation After Cardiac Injury in Mybpc^t/t Mice.

We considered whether Mybpc^t/t mice have better cardiac regenerative capacity after injury than wild-type mice, as myocyte division may play an important role in regeneration. To test this hypothesis, we performed sham or LV apical resections (26, 34), excising ∼15% of the total LV mass in P10 wild-type and Mybpc^t/t mice (Fig. 5 A–I). There is no difference in survival of the two genotypes after apical resection. Surviving neonates received a BrdU pulse 1 d later, and at P17 BrdU incorporation was assessed in hearts stained with cTnT-specific antibody, to exclude nonmyocytes, and with Ki67 antibody to detect cell cycle activity (Fig. 5 J and K). Sections from sham-operated Mybpc^t/t LV (47 ± 5 BrdU⁺;cTnT⁺ cells/mm²) had significantly more cTnT-positive, BrdU-labeled cells than sham-operated wild-type hearts (5 ± 0 BrdU⁺;cTnT⁺ cells/mm²; P = 0.005) or Ki67-labeled (18 ± 8 Ki67⁺;cTnT⁺ cells/mm² vs. 3 ± 3 Ki67⁺;cTnT⁺ cells/mm²; P = 0.05; Fig. 5 J and K). Apical resection did not increase BrdU incorporation into cTnT-positive cells from either wild-type and Mybpc^t/t mice, and Ki67 staining was not increased in either wild-type or Mybpc^t/t LV (Fig. 5 J and K).

Fig. 5.

Cell proliferation in P10 wild-type and Mybpc^t/t mice after apical resection. Sections from sham-operated or LV apical resected hearts, 17 d postresection, from wild-type (A–C) or Mybpc^t/t (D–I) mice stained with hematoxylin/eosin (A, B, and D–F) or von Kossa (C and G–I). Dashed line indicates the resection plane. Note the marked calcium precipitate in apical resected Mybpc^t/t (H and I) but not wild-type (C) hearts. [Scale bar, (A, C–E, G, and H) 1 mm and (B, F, and I) 200 µm.] Quantification of BrdU-labeled, cTnT-positive myocytes (J, BrdU⁺;cTnT⁺) and Ki67-positive, cTnT-positive myocytes (K, Ki67⁺;cTnT⁺) isolated 17 d after sham (gray) or LV apical resection (black) of wild-type and Mybpc^t/t mice. Ten sections were counted from each of four mice per genotype, and cell counts are reported as stained cells per mm² (mean ± SD).

Notably, dystrophic calcification, a marker of cell death and necrosis, was observed in apical resected regions in all Mybpc^t/t mice, but not in apical resected wild-type or sham-operated Mybpc^t/t mice (Fig. 5 B, H, and I). As cTnT-positive cells without nuclei surrounded these necrotic foci, we deduced that Mybpc^t/t myocytes were particularly sensitive to dystrophic calcification and death. The demise of Mybpc^t/t myocytes was unlikely to reflect inadequate vascular supply, as staining by the vascular endothelial marker CD31 was comparable in wild-type and Mybpc^t/t mice (Fig. S3). Dystrophic calcification also occurred in both apical and basal regions when apical resection was carried out in P1 Mybpc^t/t mice (Fig. S4). We suggest that Mybpc^t/t hearts were more susceptible to necrotic death after resection due to altered biomechanical properties that increased energy requirements due to enhanced actomyosin force (35) and/or elevated wall stress (24) in Mybpc^t/t myocytes. Mypbc^t/t mice appear to have reduced regenerative capacity compared with wild-type mice.

Fig. S3.

Cardiac, endothelial, and cell cycle markers expressed 7 d post-LV apical resection of P10 wild-type (A and B) and Mybpc^t/t mice (C and D). Heart sections were stained for DAPI or immunostained to detect cTnT (green) and Ki67 (red). An asterisk denotes location of dystrophic calcification in hematoxylin/eosin-stained Mybpc^t/t sections (Fig. S4). [Scale bars, (A and C) 200 μm; (B and D) 0.5 mm.]

Fig. S4.

Calcium precipitates in Mybpc^t/t mice after apical resection at P1. Sections from sham-operated or LV apical resected hearts, 7 d postresection, from wild-type (A and B) or Mybpc^t/t (C–H) mice stained with hematoxylin/eosin (A and C–E) or von Kossa (B and F–H). Note the marked calcium precipitate in apical resected Mybpc^t/t (G and H) but not wild-type (B) or sham (F) Mybpc^t/t hearts.

Discussion

MYBPC is critical for normal cardiac structure and function. Absence of MYBPC in the adult heart compromises the stiffness and rigidity of myofilaments (36, 37), reduces sarcomere packing density (24), and perturbs contractile performance (16), which might account for increased myocyte width in Mybpc^t/t mice and contribute to increased LV wall thickness (Fig. S1 C and D). We show that MYBPC also has unexpected roles in postnatal maturation of myocytes. Postnatal Mybpc^t/t myocytes appear morphologically immature, based on dimensions and a single nucleus, unlike postnatal wild-type myocytes. During neonatal development, Mybpc^t/t myocytes undergo additional rounds of division, resulting in increased myocyte numbers in Mybpc^t/t hearts (Fig. 6). Given its established role in myofilament lattice rigidity (1, 7, 12, 17, 18, 24), we conclude that MYBPC is critical for inhibiting postnatal myocyte cytokinesis, an essential step in myocyte maturation accompanying cell cycle exit.

Fig. 6.

A schematic of postnatal myocyte development in Mybpc^t/t and wild-type mice. In the perinatal period, wild-type myocytes develop dense mature myofibrillar structures as they undergo a final round of DNA replication without cytokinesis, resulting in 95% binuclear myocytes. Mybpc^t/t myocytes have reduced myofibrillar density and less rigid sarcomere, prolonged expression of cell cycle markers, resulting in more myocytes and larger proportions of mononuclear myocytes.

To progress through the cell cycle, myocytes must disassemble both the cytoskeleton and the contractile apparatus, the thick and thin myofilaments of the sarcomere. High-resolution confocal microscopy of immunostained neonatal myocytes has identified two phases of sarcomere dissolution. Before metaphase, proteins in the z disk and thin filament disassemble, whereas thick filament proteins, including myosin and MYBPC, maintain a mature, cross-striated pattern. Late in anaphase, thick filament proteins disassemble and remain dispersed until cytokinesis is complete (38).

Before cell cycle exit, neonatal mouse myocytes undergo karyokinesis without cytokinesis (39, 40). This phase in myocyte maturation is characterized by synchronized changes in the expression of cell cycle molecules; levels of activators (Cdk2, Cdk3, Cdk4, Cdk cofactors, and Ccnd1) are decreased, and levels of inhibitors (p21, p27KIP1, TSC2, p130, and Rb) are increased. Transgenic overexpression or genetic depletion in the heart of some cell cycle regulators (27, 41) has been demonstrated to partially overcome neonatal myocyte cell cycle arrest and enable additional rounds of cell division. Recent studies demonstrate that these reciprocal events are mediated in part by the transcriptional activator 14-3-3ε (42), the transcription factor Meis1 (43), and activation of the DNA damage response pathway induced by reactive oxygen species in neonatal mice (44).

MYBPC deficiency also impacted the regenerative capacity of the P10 mouse heart. Apical resection did not increase BrdU incorporation or Ki67 staining among cTnT-positive cells in Mybpc^t/t mice as it did in wild-type mice (Fig. 5 and Fig. S3). Moreover, as resected Mybpc^t/t hearts showed necrosis and dystrophic calcification, we suspect that MYBPC-deficient myofilaments are less resistant to sheer forces and perhaps more susceptible to depolymerization, as has been seen ex vivo (36, 45). These observations hint at the dichotomy inherent in a rigid sarcomere, a structure that can support the stress of contraction but that also impedes myocyte regeneration.

The findings presented here suggest two independent mechanisms by which MYBPC3 mutations can alter cardiac morphology. Individuals carrying heterozygous MYBPC3 mutations develop HCM (46), characterized by cardiac hypertrophy (increased heart mass and altered morphology). These changes in size and shape are due, at least in part, to the increased size of myocytes (hypertrophy) (47). By contrast, individuals homozygous for MYBPC3 mutations develop cardiac dilation, which we conclude is due to myocyte hyperplasia (increased numbers of myocytes). This myocyte hyperplasia is associated with increased numbers of mononuclear myocytes that may have reduced contractile function. Defining the relationship of homozygous MYBPC3 DCM to other more common forms of DCM will provide further insights into this clinically important pathophysiology.

Materials and Methods

Mouse Studies.

All mouse studies were performed with approved protocols in compliance with the Association for the Assessment and Accreditation of Laboratory Animal Care and Harvard Medical School. Mybpc^t/t (16) and wild-type mice (129SvEv background) were studied. The Mybpc^t/t alleles contain a PGK-neomycin resistance gene that disrupts exon 30 and is predicted to encode from a truncated peptide that terminates with amino acid residue 1,064 of the 1,270 residues in wild-type cardiac MYBPC. Protein chemical studies have demonstrated that this allele produces less than 10% of the normal amount of cardiac MYBPC.

Myocyte Isolation.

Ventricular myocytes were isolated from P1 to P10 mice using a modified collagenase dissociation protocol. After anesthetizing mice [2% (vol/vol) isoflurane inhalation], the thoracic cavity was opened and LV was perfused (2 mL/min) with prewarmed 37 °C buffer (126 mM NaCl, 4.4 mM KCl, 1 mM MgCl₂, 4 mM NaHCO₃, 30 mM 2,3-butanedione monoxime, 10 mM Hepes, 11 mM Glucose, 0.5 mM EDTA) with 0.09% Collagenase Type I (Worthington), 0.125% Trypsin, and 25 µM CaCl₂ for 2 min. The perfused heart was then excised and placed in buffer with 100 µM CaCl₂ and 2% BSA for 10 min at 37 °C. Ventricular tissue was minced, and myocytes were dispersed by gentle trituration of minced tissue through a wide bore disposable serologic pipette. The dispersed cells were filtered through a 100 µm nylon mesh and washed twice by centrifugation (50 × g for 3 min). The resulting cell pellet was suspended in 1 mL buffer with 100 µM CaCl₂ and fixed by the addition of 1 mL 2% (wt/vol) paraformaldehyde in PBS and incubated on ice for 20 min. Fixed myocytes were washed twice in PBS by centrifugation (50 × g for 3 min) and stored at 4 °C.

For ventricular myocyte isolation from 3–4-wk-old mice, immediately after sacrifice, hearts were excised, the ascending aorta was quickly cannulated, and the hearts were perfused for 5 min with warm (37 °C) calcium Tyrode’s buffer (135 mM NaCl, 4 mM KCl, 0.33 mM NaH₂PO₄, 1.2 mM MgSO₄, and 10 mM Hepes, pH 7.40), followed by 10–15 min of perfusion with collagenase B (0.56 mg/mL, Roche), collagenase D (0.48 mg/mL, Roche), and protease XIV (0.07 mg/ml, Sigma). The hearts were minced and filtered as described above. Cells were plated onto laminin-precoated coverslips (1 μg/cm²; Invitrogen) and cultured for 1 h in MEM (Sigma) containing 10 mM 2,3-butanedione monoxime to extract myocytes from other cardiac cells.

Immunohistochemistry.

Histochemical analyses were performed on heart sections fixed in 4% (wt/vol) paraformaldehyde overnight. Sections were treated with xylene (to remove paraffin), rehydrated, and permeabilized in 0.1% (vol/vol) Triton-X100 in PBS. Sections were incubated with primary antibodies applied at 1:200 dilution (unless otherwise indicated) in 0.1% (wt/vol) BSA in PBS overnight at 4 °C, and nonspecific antibody binding was blocked by 1.5% (vol/vol) FCS in PBS. Primary antibodies included the following: BrdU (rat anti-BrdU, Abcam ab6326, 1:200), Ki67 (rabbit anti-Ki67, Abcam ab15580, 1:200), cTnT (mouse anti-cTnT, Abcam ab8295, 1:200), cardiac troponin-I (rabbit anti-Tnni3, Abcam ab56357, 1:200), WGA (Invitrogen W32466, 1:400), pH3 (rabbit anti-pH3, Millipore 06–570, 1:250), and Aurora B (rabbit anti-Aurora B, Abcam ab2254, 1:250). Sections were washed in PBS and fluorophore-conjugated secondary antibody (Molecular Probes) diluted 1:200 in 1% FCS or donkey serum. Nuclei were counterstained with 0.3 µg/mL propidium iodide (Molecular Probes) or 2 µg/mL of 4’,6-diamidino-2-phenylindole (DAPI; Sigma). Myocytes were identified by visualizing sarcomeres with Nomarski (differential interference contrast) light microscope optics and cTnT expression. Fluorescent images of histological sections were captured using the Leica TCS NT confocal microscope system and subsequently analyzed and assembled using Image J and Adobe Photoshop software.

Myocyte Quantification.

WGA-stained heart sections were imaged using a Zeiss confocal microscope, and images were processed after color inversion, so that individual cells were highlighted as “particles.” The “Watershed” algorithm (Fiji; fiji.sc/) was used to resolve distinct particles that were close together. Particles were included based on the size of cardiomyocytes ranging from 150 to 3,000 pixels/cell.

AAV Production and Purification.

AAV vectors were packaged into AAV9 capsid by the triple transfection method using helper plasmids pAdΔF6 and plasmid pAAV2/9 (Penn Vector Core). We used 50 μg of plasmid DNA per 15-cm cell culture plate. Three days after transfection, AAV vectors were purified in Optiprep density gradient medium (d-1556, Sigma) by centrifugation and stored at −80 °C.

shRNA Vector Construction.

Two shRNA constructs specific for 21-base-pair sequences corresponding to cardiac Mypbc3 were constructed. The following targeted sequences were used: Mypbc3-specific shRNA-1, ccagagaaggcagaatctgaa; Mypbc3-specific shRNA-2, aagggtttgcctgcaacctgt; and shRNA control targeting LacZ, gactacacaaatcagcgattt.

RNA-Seq.

Hearts from mice were rapidly isolated, placed in room temperature PBS to evacuate blood, and then immersed in RNALater (Qiagen) at room temperature. We used 2 μg of total ventricular RNA to construct RNAseq sequencing libraries (48). At least 20 million 50-bp paired-end DNA reads were obtained from each library using the Illumina HiSeq2500. Data were processed as previously described (48).

Apical Resection.

LV apical resection procedures were performed as described (26). In brief, mice were anesthetized on ice to induce deep transient sedation, apnea, and asystole. The thoracic cavity was opened at the fourth intercostal space and the LV exposed. Approximately 15% of the ventricular apex was resected. Successful apical resection was assessed by visualization of the ventricular chamber immediately after resection. Following surgery, mice were rapidly warmed and monitored for viability. Sham-operated mice underwent identical procedures but without LV apex resection. We have independently verified that robust heart regeneration occurs in the neonatal mouse apical resection model (34).

EdU/BrdU Labeling.

To assess DNA synthesis in isolated myocytes, neonatal pups received an i.p. injection of EdU (Life Technologies, C10339, 5 mg/kg) for the first 5 d. To label cells with ongoing DNA synthesis, surviving mice subjected to sham or apical-resection surgery received a pulse (two s.c. injections 12 h apart) of BrdU (Sigma, 20 mg/mL solution in sterile PBS, 100 mg/kg dose) on the first day post-resection.

Statistical Analyses.

Significance was assessed by two-sample Student’s t test on selected groups, assuming two-tail heteroscedastic distributions. Multiple group comparison was assessed by ANOVA.